I am trying to determine allele frequency of a hla allele. Preliminary analysis shows the prevalence of the allele (homo or heterozygous both) is 3%. The allele is related to a disease (After exposure to risk factor). Taking samples from college or university can help reduce time and money. Taking samples from hospital may take time (the hospital I have contact to, has small sample size?). Though both ways may have ethnic diversity, college students are of a given generation only. What to do. Please suggest with references to read.
I suggest this as a starting point.
Basically, if your aim is to determine the population frequency of the alleles, then you should sample from the population. A possible problem with the university is the age stratification, but also possible different proportion of ethnic groups compared to general population. Also hospital might be biased, because you will have a higher prevalence of people with health issues. The best thing is to randomly collect samples from the population (but I know it is very hard, especially because you can sample randomly, but the ratio of people willing to give you their blood will be biased!).
One thing that is commonly accepted is to retrieve data from healthy blood donors (which is not unbiased, but is accepted!). You can give a look at a paper I published a lot of time ago. In that case, the estimation of haplotype frequency was performed on the bone marrow donors registry, and it was considered perfectly fine.